Istio sidecar injection

There are several ways to inject istio sidecar configuration into Pods. For example: automated injection, YAML/JSON deployment update, using Helm or Kustomize and update of existing live deployment. We will look into each of them.

Automatic Sidecar injection

Istio uses ValidatingAdmissionWebhooks for validating Istio configuration and MutatingAdmissionWebhooks for automatically injecting the sidecar proxy into user pods.

For automatic side car injection to work admissionregistration.k8s.io/v1beta1 should be enabled:

$ kubectl api-versions | grep admissionregistration.k8s.io/v1beta1
admissionregistration.k8s.io/v1beta1

Step two is to verify MutatingAdmissionWebhook and ValidatingAdmissionWebhook plugins are listed in the kube-apiserver –enable-admission-plugins. That can be done by cluster administrators.

When the injection webhook is enabled, any new pods that are created will automatically have a sidecar added to them.

To enable namespace for sidecar injection label the namespace with istio-injection=enabled

$ kubectl label namespace default istio-injection=enabled
$ kubectl get namespace -L istio-injection
NAME           STATUS    AGE       ISTIO-INJECTION
default        Active    1h        enabled
istio-system   Active    1h
kube-public    Active    1h
kube-system    Active    1h

Sidecar injection with istioctl on YAML file

To manually inject side into deployment, use istioctl kube-inject

$ istioctl kube-inject -f deployment.yaml | kubectl apply -f -

Sidecar injection into existing deployment

$ kubectl get deployment -o yaml | istioctl kube-inject -f - | kubectl apply -f -

Sidecar injection with istioctl and helm

Sidecar injection into helm release could be done in two steps. We will use helm install and helm template to inject sidecar. As down side some features as rollback of helm release wouldn’t work, only rolling forward would be possible.

First. Using helm install we install the package:

$ helm install nginx stable/nginx

Step two update the deployment with sidecar using helm template:

$ helm template stable/nginx | istioctl kube-inject -f - | kubectl apply -f -

Sidecar injection with kustomize

Deployment file:

resources:
- deployments.yaml

To inject istio sidecar into deployment Kustomize patch should be used

patches:
- path: sidcar.yaml
  target:
    kind: Deployment

Where sidecar.yaml is istio sidecar deployment.

Conclusions

There are many ways to install istio sidecar or any sidecar into deployment. The main idea is to render deployment file and wrap it up with istioctl for manual injection or setup automatic injection with Admission webhook.

How to organize Namespaces in Kubernetes

There are two main objectives:

  1. Users are able to do their job with the highest velocity possible
  2. Users organized by groups in multi tenant setup 

Multi tenancy

Kubernetes namespaces help to setup boundaries between groups of users and applications in a cluster.
To make it more pleasant and secure for your users to work in shared cluster Kubernetes has a number of policies and controls.

Access policies

RBAC primary objective is authorize users and applications to do specific operations in the namespace or in whole cluster. Use RBAC to give your users enough permissions in the namespace, so they can do day to day operations on their own.
Network Policy control how pods can communicate with each other. Use it to firewall traffic between namespaces or inside namespace to critical components like Databases.

Resource controls

By default Pod can utilize as many compute resources as available.
Resource Quotas control the amount of compute and storage resources which Pod can use in namespace.
Limit Range help to prevent one Pod from utilize of all resources in namespace. LimitRange set minimum and maximum boundaries for compute and storage resource per Pod.

Application Security

Pod security policy control security sensitive aspects of container. Examples are privileged containers, use of host namespace and many other.
Open Policy Agent is very powerful policy framework which help to create custom policies for applications and users in a cluster. For example:

  • force users to use a specific label in Kubernetes objects like Service or Deployment
  • deny access to pull :latest images tag
  • allow to pull images only from specific docker registry

Namespaces


Following examples could help you to decide on namespaces boundaries and naming:

  • Namespace per team
  • Namespace per team and project
  • Namespace per application
  • Namespace per git branch name

Namespace should provide enough self managing autonomy for users and be in sync with applications requirements.
The bigger namespace the harder to tune up it’s boundaries, at the same time many small namespaces could create additional operational work for cluster administrators.

Namespace per team and project is optimal start which should work for most organizations.

Let me know your experience in comments and have a great day!

120 Days of AWS EKS in Staging

Felix Georgii wakeboarding at Wake Crane Project in Pula, Croatia on September 25, 2016

My journey with Kubernetes started with Google Kubernetes Engine then one year later with self managed kuberntes and then with migration to Amazon EKS.

EKS as a managed kubernetes cluster is not 100% managed. Core tools didn’t work as expcted. Customers expectation was not aligned with functions provided. Here I have summarized all our experience we gained by running EKS cluster in Staging.

To run EKS you still have to:

  • Prepare network layer: VPC, subnets, firewalls…
  • Install worker nodes
  • Periodically apply security patches on workers nodes
  • Monitor worker nodes health by install node problem detector and monitoring stack
  • Setup security groups and NACLs
  • and more

EKS Staging how to?

EKS Setup

  • Use terraform EKS module or eksctl to make installation and maintenance easier.

EKS Essentials

  • Install node problem detector to monitor for unforeseen kernel or docker issues
  • Scale up kube-dns to two or more instances
  • See more EKS core tips in 90 Days EKS in Production

EKS Autoscaling

  • Kubernetes cluster autoscaling is no doubt must have addition to EKS toolkit. Scale your cluster up and down to 0 instances if you wish. Base your scaling on cluster state of Pending/Running pods to get maximum from it.
  • Kubernetes custom metrics, node exporter and kube state metrics is must have to enable horizonal pod autoscaling based on build in metrics like cpu/memory and as well on application specific metrics like request rate or data throughput.
  • Prometheus and cadvisor is another addition you would need to enable metrics collection

Ingress controller

  • Istio one of the most advanced, but breaking changes and beta status might introduce hard to debug bugs
  • Contour looks like good replacement to Istio. It didn’t have that good community support as istio, but stable enough and has quite cool CRD IngressRoute which makes Ingress fun to use
  • Nginx ingress is battle tested and has the best support from community. Have huge number of features, so is a good choice to setup the most stable environment

Statefull applications

  • Ensure you have enough nodes in each AZ where data volumes are. Good start is to create dedicated node group for each AZ with minimum number of nodes needed.
  • Ensure persistent volume claim(PVC) is created in desired AZ. Create dedicated storage class for specific AZ you need PVC to be in. See allowedTopologies in following example.
kind: StorageClass
apiVersion: storage.k8s.io/v1
metadata:
  name: standard-eu-west1-a
provisioner: kubernetes.io/aws-ebs
parameters:
  type: gp2
volumeBindingMode: WaitForFirstConsumer
allowedTopologies:
- matchLabelExpressions:
  - key: failure-domain.beta.kubernetes.io/zone
    values:
    - eu-west1-a

Summary

EKS is a good managed Kubernetes service. Some of mentioned tasks are common for all Kubernetes platforms, but there is a lot of space to grow for the better service. The burden for maintenance is still quite high, but fortunately Kubernetes ecosystem has a lot of opensource tools to easy it.

Have fun!

Kubernetes sidecar pattern: nginx ssl proxy for nodejs

I learn about sidecar pattern from Kubernetes documentation and later from blog post by Brendan Burns The distributed system toolkit. Sidecar is very useful pattern and work nice with Kubernetes.
In the tutorial I want to demonstrate how “legacy” application can be extend with https support by using  sidecar pattern based on Kubernetes.

Problem

We have legacy application which doesn’t have HTTPS support. We also don’t want to send plain text traffic over network. We don’t want to make any changes to legacy application, but good thing that it is containerised.

Solution

We will use sidecar pattern to add HTTPS support to “legacy” application.

Overview

Main application
For our example main application I will use Nodejs Hello World service (beh01der/web-service-dockerized-example)
Sidecar container 
To add https support I will use Nginx ssl proxy (ployst/nginx-ssl-proxy) container

Deployment

TLS/SSL keys
First we need to generate TLS certificate keys and add them to Kubernetes secrets. For that I am using script from nginx ssl proxy repository which combine all steps in one:
git clone https://github.com/ployst/docker-nginx-ssl-proxy.git
cd docker-nginx-ssl-proxy
./setup-certs.sh /path/to/certs/folder

Adding TLS files to Kubernetes secrets

cd /path/to/certs/folder
kubectl create secret generic ssl-key-secret --from-file=proxykey=proxykey --from-file=proxycert=proxycert --from-file=dhparam=dhparam

Kubernetes sidecar deployment

In following configuration I have defined main application container “nodejs-hello” and nginx container “nginx”. Both containers run in the same pod and share pod resources, so in that way implementing sidecar pattern. One thing you want to modify is hostname, I am using not existing hostname appname.example.com for this example.
apiVersion: apps/v1beta2
kind: Deployment
metadata:
  name: nodejs-hello
  labels:
    app: nodejs
    proxy: nginx
spec:
  replicas: 1
  selector:
    matchLabels:
      app: nodejs-hello
  template:
    metadata:
      labels:
        app: nodejs-hello
    spec:
      containers:
      - name: nodejs-hello
        image: beh01der/web-service-dockerized-example
        ports:
        - containerPort: 3000
      - name: nginx
        image: ployst/nginx-ssl-proxy
        env:
        - name: SERVER_NAME
          value: "appname.example.com"
        - name: ENABLE_SSL
          value: "true"
        - name: TARGET_SERVICE
          value: "localhost:3000"
        volumeMounts:
          - name: ssl-keys
            readOnly: true
            mountPath: "/etc/secrets"          
        ports:
        - containerPort: 80
          containerPort: 443
      volumes:
      - name: ssl-keys
        secret:
          secretName: ssl-key-secret

Save this file to deployment.yaml and create deployment Kubernetes object:

kubectl create -f deployment.yaml

Wait for pods to be Read:

kubectl get pods

NAME                            READY     STATUS    RESTARTS   AGE
nodejs-hello-686bbff8d7-42mcn   2/2       Running   0          1m

Testing

For testing I setup two port forwarding rules. First is for application port and second for nginx HTTPS port:

kubectl -n test port-forward <pod> 8043:443
#and in new terminal window run
kubectl -n test port-forward <pod> 8030:3000

First lets validate that application respond on http and doesn’t respond on https requests

#using http
curl -k -H "Host: appname.example.com" http://127.0.0.1:8030/ 
Hello World! 
I am undefined!

#now using https
curl -k -H "Host: appname.example.com" https://127.0.0.1:8030/ 
curl: (35) Server aborted the SSL handshake

Note: SSL handshake issue is expected as our “legacy” application doesn’t support https and even if it would it must serve https connection on different port than http. The test goal was to demonstrate the response.

Time to test connection through sidecar nginx ssl proxy

curl -k  -H "Host: appname.example.com" https://127.0.0.1:8043/
Hello World!
I am undefined!

Great! We have got expected output through https connection.

Conclusions

  • Nginx extended nodejs app with https support with zero changes to any of containers
  • Sidecar pattern modular structure provide great re-use of containers, so teams can be focused on application development
  • Ownership of containers can be split between teams as there is no dependency between containers
  • Scaling might not be very efficient, because sidecar container have to scale with main container